Real-time processing
NXT AP embeds Azure AI-services which provide real-time processing. Other than traditional processes with turnaround cycles of hours or even days, NXT AP receives complete supplier invoice data in less than 2 seconds.
All supplier invoices are parked within less than 5 seconds after receipt (PDF or e-invoice data), which gives real-time visibility of the financial standing of the company with view of its payables Parked documents in SAP can be leveraged for various reporting and planning purposes beyond the reports provided by NXT AP itself. By leveraging these standard SAP reports in addition to NXT AP's own reports, corporate customers can gain a comprehensive view of their obligations and cash flow, enabling more informed decision-making and planning. Some of the standard SAP reports that can be useful for corporate customers in respect to obligation and cash flow planning include:
Open Items Report (FBL1N/FBL5N): This report shows all open customer or vendor items, including parked documents, which can be helpful for obligation planning.
Cash Flow Statement (S_ALR_87012352): This report provides an overview of cash inflows and outflows, including those related to parked vendor invoices, which can aid in cash flow planning.
Liquidity Forecast (RFUCOBL0): This report forecasts a company's future liquidity position based on open items, including parked vendor documents, which is crucial for cash flow management.
Vendor Aging Analysis (FBL5N): This report shows the aging of open vendor items, including parked invoices, which can help identify upcoming payment obligations.
Vendor Payment History (FBL1N): This report provides a detailed history of vendor payments, including those made against parked documents, which can be useful for analyzing payment patterns and cash flow.
Stateless processing
NXT AP embeds Azure AI services to capture supplier invoice data from image-based invoices. Azure does not keep any of the captured data, all processing is stateless. This ensures data security standards.
Also, NXT AP embeds Azure AI services for complex data validation, e.g., matching payment terms or line-item descriptions etc. None of this data is used for further training, neither by nxt4sap nor by Azure. This data is exclusively used in the customer-specific subscription.
Master data less processing
NXT AP embeds Azure AI services to capture supplier invoice data from image-based invoices. In contrast to traditional capture services, NXT AP never provides any supplier master data nor PO data or any other data to support the capture services.
Confidence score
Azure Ai gives a confidence score > 95% often, but not always regardless how derived, 95% leaves 5% of doubt - per value, across all documents.
Financial accounting is not about "confidence".
We are pretty certain that this is the invoice number is not good enough in an environment where thousands of invoices are automatically processed (STP).
NXT AP validates every single value, 100%, in real-time
Matching
NXT AP matches supplier invoice data with SAP data to make the data useful for AP processing. If matching is done "outside", it requires the data used to be copied out. Regularly, the more often the better. If matching goes beyond simple supplier ID and PO line items, but includes material data, data related to GR/SES entries, payment terms, etc. the data required for matching is huge
NXT AP matches the supplier invoice data with the COMPLETE and CURRENT data of the entire SAP system
Full text
Supplier invoice data is categorized as "semi-structured". Most values are expected in some structure, e.g., amounts, dates, etc. Lot's of data is not, and it never ensured, even for electronic invoices, often needed information, such as material specifications or delivery note references are on the full-text. Same goes for payment terms or assignments, which are "hidden" in the "fineprint".
Azure AI's standard model for supplier invoices provides more than 50 fields on header level and more than 10 on line-item level, but without full-text analysis many validations will fail
NXT AP keeps the full-text available until supplier invoice are posted, allowing to match values even after the supplier invoice data has already been processed. Example: Invoices with delivery note references received before the GR is posted.
Integrating with national e-Invoice networks natively
More and more markets around the world provide national e-Invoice networks which are governed by national authorities and are thus by definition compliant with local regulations. In increasing number of markets even mandate the use of these e-Invoice networks.
By leveraging the direct integration with these networks, NXT AP can access complete supplier invoice data without any data capturing efforts.
In its core, NXT AP provides a structured processing sequence for supplier invoice data, independent whether it originates from national e-invoice networks or is captured from PDF documents using Azure AI services to match and validate the data and eventually allow a reliable and compliant straight-through-processing from receipt to posting.
NXT AP ensures that all supplier invoices, regardless of their original format, are processed identically, using the identical set of matching, validation, and processing rules.
Neither business users nor accountants need to differentiate supplier invoices based on their format or data source.
Validation
NXT AP provides comprehensive validations, such as whether the invoice is compliant with applicable regulatory requirements", whether the "Invoice vendor" is the "PO vendor" or the "Condition vendor", whether the supplier is a "Non-PO" vendor, whether the invoice number captured is actually the right value.
NXT AP's set of validations grows with experience and project requirements. Validations are the core of NXT AP. Our list of validations are already very complete because NXT AP complements the already existing set of existing SAP validations and because of our teams 20 years of experience especially in global high-volume shared services centers.
NXT AP processes are designed based on a configurable and flexible sequence of validations from receipt to posting.
Accounting rules
Nothing in financial accounting is "by opinion", everything is rule-based: Laws, regulatory requirements, accounting standards, and agreed and documented processes determine how financial transactions are eventually documented.
While in high-volume environments most supplier invoices are based on purchase orders which will address most aspects of these accounting rules (assignment of cost objects, such as costs centers & GL accounts, etc.), supplier invoice transactions which are not based on purchase orders ("FI invoices") require rules-based coding per line.
Competitors promote machine-learning based coding proposals, which is the equivalent of "we have always done it this way" and "this coding will be correct - with a high-degree of certainty", both arguments will certainly lead to questions by auditors and authorities.
NXT AP provides AI-based coding proposals for GL postings (non PO-lines) which are based on GPT and its ability to "learn" and "understand" accounting rules, if GPT does not already "know" these rules as part of its comprehensive training (we are using Azure AI's current GPT-4 model)
GL line coding proposals are not based on statistical probability but on comprehensive understanding of accounting rules and access to all relevant data (complete supplier invoice data, complete SAP data, complete documentation of regulatory and other accounting requirements and standards. NXT AI is simply the best trained, most experienced accounts payable accountant, available 24/7, across language barriers and for all markets.
GPT for superhuman validation
Certain validations require "language capabilities" that only human accountants and GPT have. If lines item descriptions are to be considered, payment-terms are to be identified and matched, or a seemingly endless lists of random material codes (e.g., HSN numbers ...), only GPT can identify, match and validate. 24/7 and across all languages
The SAP-perspective
Without doubt, by far most financial transactions world-wide are recorded in SAP systems. The objective of NXT AP for SAP is to ensure that supplier invoice data can be processed AUTONOMOUSLY, that straight-through-processing (STP) is RELIANT, EFFICIENT and COMPLIANT.
NXT AP embeds most advanced AI-technology from Computer Vision to GPT4.
For those new to the series, this blog post is part of a broader series aimed at highlighting the unique capabilities of NXT AP compared to traditional solutions in the market. Each post explores a different aspect of our innovative accounts payable solution.
Contact us today to learn more about NXT AP's innovative approach to AP automation and to schedule a live demonstration.
_______________________________________________________________________________________________________
At NXT4SAP, we fuse GPT technology, process expertise, and decades of experience to empower global clients in making strategic decisions. Our adaptive software solutions cater to Asian requirements, especially in finance automation for corporates and Shared Services Centres.
The team excels in AI-driven business process automation, strategy, and innovation.
Our global experience with Fortune 500 companies, particularly in SE Asia, enhances our understanding of diverse requirements. While specializing in finance processes with SAP, we prioritize holistic end-to-end solutions.
Reach us at info@nxt4sap.com to redefine your global business dynamics through AI-driven solutions.
Comments